1

Artificial Intelligence
An Internship Report

Submitted in the partial fulfillment of requirements to

Summer Internship (CS- 451)

IV/IV B. Tech (VII Semester)

Submitted By

Manyam Abhiram (Y21CS095)

November, 2024

R.V.R. & J.C. COLLEGE OF ENGINEERING (Autonomous)

(Affiliated to Acharya Nagarjuna University)
Chandramoulipuram :: Chowdavaram,
GUNTUR – 522019, Andhra Pradesh, India
 2

Internship Certificate
 3

ACKNOWLEDGEMENT

The successful completion of any task would be incomplete without proper suggestions,
guidance and environment. The combination of these three factors acts like backbone to our “Artificial
Intelligence” Internship

We would like to express our profound gratitude to Dr. M. Sreelatha, Head of the Department
of Computer Science and Engineering and Dr. Kolla Srinivas, Principal of R.V.R. & J.C. College of
Engineering, Guntur for their encouragement and support to carry out this Internship successfully.

We are very much thankful to at AIMER Society - Artificial Intelligence Medical and
Engineering Researchers Society, for having allowed us to conduct the study needed for the Internship.

Finally, we submit our heartfelt thanks to all the staff in the Department of Computer Science
and Engineering and to all our friends for their cooperation during the Internship work.

Manyam Abhiram(Y21CS095)
 4

ABSTRACT

I am excited to express my interest in the Artificial Intelligence internship opportunity at

AIMER Society. As a Computer Science student, I am eager to apply and further develop my skills in
a practical, hands-on environment. With a strong foundation in Python and data analysis, I am passionate
about creating innovative AI solutions that foster intelligent automation and enhance decision making
processes.

During my previous internship, I had the opportunity to apply theoretical knowledge in real-
world scenarios. I independently developed machine learning models and algorithms, translating
complex AI concepts into practical applications. This experience not only honed my proficiency in AI
frameworks but also strengthened my understanding of model optimization and data processing.
Through this, I cultivated a data-driven mindset and enhanced my problem-solving skills, which I
believe are essential for succeeding in the rapidly evolving field of Artificial Intelligence.

I am particularly excited about the opportunity to develop AI-powered models using machine
learning libraries and frameworks. Whether it involves predictive modeling, natural language
processing, or computer vision, I am eager to contribute to impactful projects while further advancing
my technical skills.

I am confident that my technical skills, coupled with my dedication to learning and

collaboration, make me a strong candidate for the Artificial Intelligence internship at AIMER Society.
 1

Contents

Page No.
Title Page i
Certificate ii
Acknowledgement iii
Abstract iv
Contents
1. Introduction 1
1.1 Background 2
1.2 Important terms in Artificial Intelligence 3
1.3 Objectives 4
2. SYSTEM ANALYSIS
2.1 Requirements 5
2.1.1 Functional Requirements 6
2.1.2 Non Functional Requirements 7
3. System Design
3.1 Architecture Diagram 8
3.2 Workflow 11
3.3 Module Description 12
4. Implementation 14
5. Result Analysis 28
6. Conclusion and Future Work 31
 2

1. Introduction

1.1 Background
Artificial Intelligence (AI) is a transformative technology designed to enhance decision making,
automate processes, and analyze vast amounts of data. Businesses typically employ AI professionals to
develop intelligent systems that optimize operations, improve customer experiences, and drive
innovation.

As the volume of data generated online increases daily, the need for AI-driven solutions to handle
and process this data efficiently is growing. AI technologies can help address the rising complexity of
challenges, from automating data analysis to detecting patterns in real-time. Historically, organizations
have relied on manual or rule-based systems to handle tasks, but with the integration of AI, there’s a shift
towards proactive, adaptive, and scalable solutions. The use of AI can significantly boost efficiency,
streamline processes, and uncover insights that traditional approaches might miss.

1.2 Important terms in Artificial Intelligence

● Machine Learning (ML) – A subset of AI that involves training algorithms to learn

patterns from data and make decisions or predictions without being explicitly
programmed.
● Neural Networks – Computational models inspired by the human brain, composed of
interconnected nodes (neurons) that process data and learn from it.
● Deep Learning (DL) – A branch of machine learning that uses neural networks with
many layers to model complex patterns in data, such as images, speech, and text.

● Natural Language Processing (NLP) – A field of AI focused on the interaction

between computers and human language, enabling machines to understand, interpret,
and respond to human language.

● Computer Vision – An AI field that enables machines to interpret and understand

visual data from the world, such as images and videos.

● Reinforcement Learning – A type of machine learning where an agent learns to make

decisions by interacting with an environment and receiving rewards or penalties based
on its actions.

● Supervised Learning: A machine learning technique where an algorithm is trained on

labeled data to predict outcomes based on input-output pairs.
 3

● Unsupervised Learning: A machine learning approach where the algorithm learns

patterns from unlabeled data, often used for clustering and dimensionality reduction.

1.3 Objectives

Object Detection using YOLOV8:

The goal of employing YOLOv8 (You Only Look Once version 8) for object detection is to
precisely identify and locate multiple objects in images or videos in real time. This entails achieving
rapid processing speeds for immediate detection and classification while ensuring high accuracy to
reduce false positives and negatives. YOLOv8 seeks to recognize and categorize various objects
simultaneously, maintaining resilience against variations in size, orientation, and occlusion.
Additionally, the model is designed for scalability and adaptability, facilitating its application across
different datasets and use cases. By integrating effectively with other technologies, YOLOv8 offers a
robust solution for industries such as security, healthcare, retail, and transportation, where dependable
object detection is critical.

Telegram ChatBot:

The objective of developing a Telegram ChatBot is to provide users with an interactive platform
for communication and information retrieval within the Telegram messaging app. This involves
automating responses to user inquiries, facilitating quick access to services, and enhancing user
engagement through personalized interactions. By leveraging natural language processing and
integration with external APIs, the ChatBot aims to deliver timely and relevant information, making it a
valuable tool for businesses looking to enhance their digital presence and customer interactions.

Visual Question Answering Model:

The goal of developing a Visual Question Answering (VQA) model is to enable machines to
understand and analyze visual content in conjunction with natural language queries. This allows users
to ask questions about images or videos and receive accurate, contextually relevant answers. The model
is trained to efficiently extract and interpret visual features while integrating them with language
comprehension to provide precise responses to a wide variety of queries. By improving the interaction
between humans and machines, VQA models hold significant potential for applications in fields such as
education, accessibility, and data retrieval, ultimately enhancing user engagement and simplifying
access to visual information.
 4

Data Visualization –Power BI:

The objective of using Power BI for data visualization is to transform complex datasets into
insightful and interactive visual representations that facilitate informed decision-making. By leveraging
Power BI's powerful analytics and visualization tools, users can easily explore data trends, patterns, and
relationships, enabling them to gain actionable insights quickly. The platform aims to simplify the
process of data analysis by providing a user-friendly interface that allows for the creation of dynamic
dashboards and reports. Ultimately, Power BI seeks to empower organizations to harness their data
effectively, driving better strategic planning and enhancing overall business performance.

Image Classification:

The purpose of image classification is to automatically assign predefined labels or categories to

images by analyzing their content. This process trains models on extensive datasets to recognize
distinctive features and patterns linked to each category, enabling accurate identification of objects,
scenes, or activities within images. Image classification plays a crucial role in various applications,
including automated image tagging, facial recognition, medical image analysis, and autonomous
vehicles, significantly improving the efficiency and accuracy of visual data processing and
interpretation.
 5

2. SYSTEM ANALYSIS

2.1 Requirements

For an Artificial Intelligence internship, the recommended system specifications include a multi-
core processor (such as Intel i5 or AMD Ryzen 5), a minimum of 16GB of RAM (with 32GB being
ideal), and at least a 512GB SSD to ensure quick data access. A modern operating system like Windows
10/11, macOS, or Linux is required. Familiarity with AI frameworks like TensorFlow or PyTorch is
important, along with experience using tools like Jupyter Notebook or Anaconda. A dedicated GPU is
highly recommended for accelerated computations, and a stable, high-speed internet connection is
essential for cloud resource access and collaborative work on projects.

2.1.1 Functional Requirements

Planning:
Objective: Define the purpose and goals of the AI-driven website or web application.
Tasks: Conduct market research to explore AI trends and potential applications, identify the target
audience's needs and preferences for AI solutions, outline key features and functionalities that leverage
AI (e.g., personalized recommendations, automated responses), create a project timeline that includes
AI development phases, and establish a budget accounting for data acquisition, model training, and
deployment costs.
 6

Design:
Objective: Develop an engaging visual and user interface (UI) design that enhances user interaction with
AI features.
Tasks: Create wireframes and prototypes emphasizing AI functionalities, design the layout to facilitate
seamless user engagement with AI tools, select appropriate color schemes and typography that reflect
the brand and enhance readability, and ensure the design is user-friendly and visually appealing while
highlighting AI capabilities.

Development:
Objective: Implement the AI algorithms and functionalities to create a fully operational web application.
Tasks: Develop the backend infrastructure to support AI model integration, create data pipelines for
input and output processing, implement machine learning algorithms for specific tasks (e.g., image
recognition, natural language processing), and ensure the application can scale effectively to handle
varying user loads and data.

Testing:
Objective: Validate the performance and reliability of the AI features within the application.
Tasks: Conduct unit testing for individual AI components, perform integration testing to ensure seamless
interaction between the AI models and the application, execute system testing to verify the overall
functionality, and assess model accuracy and performance under various scenarios. Address any
identified issues or inconsistencies in AI responses.

Delivery:
Objective: Deploy the AI-powered application to a live environment for user access.
Tasks: Select a hosting solution that accommodates AI workloads and scalability, set up the domain and
necessary server configurations, transfer the application files, and ensure that AI models are properly
deployed and operational in the production environment. Facilitate a smooth transition from the
development phase to the live application.

Maintenance:
Objective: Ensure the continued performance and relevance of the AI application post-launch.
Tasks: Regularly update the AI models to improve accuracy and functionality based on user interactions
and feedback, monitor application performance to identify and resolve any emerging issues, implement
security updates to protect user data and the AI system, and adapt the application based on evolving user
needs and technological advancements.
 7

2.1.2 Non-Functional Requirements

Performance:
Ensuring that AI models operate efficiently with minimal latency, especially in real-time applications.
Scalability:
The AI system should be capable of handling increasing volumes of data and users without
compromising performance.
Accuracy:
Maintaining high precision in model predictions and outputs to meet desired levels of reliability.
Security:
Protecting AI models and data from malicious attacks, ensuring data privacy and secure access.
Usability:
Designing user-friendly interfaces for ease of interaction with AI features and making outputs
interpretable.
Maintainability:
The AI system should be easy to update, retrain, or improve as new data becomes available.
Reliability:
Ensuring the AI system consistently produces accurate results without failure, even under varying
conditions.
Ethical Compliance:
AI should adhere to ethical standards, avoiding bias and ensuring fairness in decision-making.
 8

3. System Design
3.1 Architecture Diagram

YoloV8 Architecture:
YOLOv8 is an improved version of the YOLO object detection model, designed for realtime use.
It starts with a convolutional neural network (CNN) backbone to extract features from images. Then, a
neck structure (like a Feature Pyramid Network or Path Aggregation Network) processes these features
to better detect objects at various sizes. YOLOv8 uses an advanced head to predict the locations of
objects, their classes, and how likely they are to be present. It improves both speed and accuracy by
using anchor-free detection and better loss functions, making it highly efficient for real-time object
detection tasks.
 9

Fig 3.1.1 YoloV8 Architecture

 10

Power BI Architecture:

Power BI architecture is built around several key components. First, it connects to various data sources
like databases, cloud platforms, and Excel files using Power Query, which handles data extraction,
transformation, and loading (ETL). The transformed data can either be stored in Power BI’s in-memory
model for fast performance or accessed in real-time using Direct Query. Users design data models and
create visualizations in Power BI Desktop, which can include relationships, calculated measures, and
interactive reports. These reports are published to the Power BI Service, where they can be viewed on
the web or mobile devices and shared with others. Power BI Gateway allows secure, realtime access to
on-premise data sources, ensuring up-to-date information.

Fig 3.1.2 Power BI Architecture

 11

Telegram ChatBot Architecture

A Telegram ChatBot’s architecture consists of key parts that manage user communication and
response generation. When a user sends a message through the Telegram app, it reaches the bot using
the Telegram Bot API, either through webhooks or long polling. A message handler processes the
incoming text, often using Natural Language Processing (NLP) to understand the user’s intent. The bot
then replies with predefined responses, dynamic data from external APIs, or AI-generated content.
Additionally, a database may store user data to tailor responses, ensuring smooth and personalized
interactions.

Fig 3.1.3 Telegram ChatBot Architecture

3.2 Workflow
3.2.A Workflow of Power BI Architecture
The Power BI architecture workflow begins by connecting to various data sources, such as
databases, cloud services, or flat files. The data is then extracted, transformed, and loaded (ETL) using
Power Query for cleaning and shaping. Afterward, the processed data is stored in Power BI’s inmemory
model or accessed via Direct Query for real-time analysis. Users then create reports and dashboards by
visualizing the data in Power BI Desktop, defining relationships, and performing calculations. Once
reports are finalized, they are published to the Power BI Service, where they can be shared, accessed via
 12

mobile or web, and integrated with collaboration tools. Regular data refreshes and secure access to on-
premise data can be managed through Power BI Gateway.

3.2.B Workflow of YoloV8 Architecture

The YOLOv8 architecture workflow begins by resizing the input image, which is then passed
through a convolutional backbone to extract key visual features. These features are fed into the neck,
often using a Feature Pyramid Network (FPN) or Path Aggregation Network (PAN), which enhances
detection at different scales. The detection head then processes this information to predict bounding
boxes, object classes, and confidence scores. YOLOv8's workflow is optimized for real-time object
detection, using anchor-free methods and improved loss functions to ensure fast, accurate predictions
across varying object sizes and positions.

3.2.C Workflow of Telegram ChatBot Architecture

The workflow of a Telegram ChatBot architecture begins when a user sends a message or
command through the Telegram app, which is routed to Telegram's servers. The message is then
forwarded to the bot via the Telegram Bot API, enabling it to interact with user inputs. The bot employs
either long polling or a webhook to receive these messages continuously. Upon receiving a message, the
bot processes it through a message handler, which may utilize Natural Language Processing (NLP) to
understand user intent. Based on this analysis, the bot generates a response using predefined answers,
external API calls, or dynamic content, and sends the reply back to the user through the Telegram API.
Optionally, user interactions and preferences may be stored in a database to personalize future
interactions, completing the workflow.

3.3 Module Description

Data Visualization – Power BI

● Data visualization is the representation of data through use of common graphics, such as charts,
plots, infographics and even animations.
● These visual displays of information communicate complex data relationships and data-driven
insights in a way that is easy to understand.
● Data visualization can be utilized for a variety of purposes, and it’s important to note that is not
only reserved for use by data teams.
● Data and information visualization is the practice of designing and creating easy-tocommunicate
and easy-to-understand graphic or visual representations of a large amount of complex
quantitative and qualitative data and information with the help of static, dynamic or interactive
visual items.
● Power BI is a data visualization tool developed by Microsoft.
● It is designed to help users create interactive and shareable reports and dashboards.
 13

● Power BI enables users to connect to multiple data sources, transform and clean the data, and
then visualize it using various types of charts, graphs, and maps.
● It is widely used for business intelligence and data analysis purposes, allowing users to gain
insights and make data-driven decisions.

Object Detection - YoloV8

The object detection module in YOLOv8 is designed for real-time, efficient object localization and
classification. It consists of the following key components:

1. Backbone: A Convolutional Neural Network (CNN) that processes the input image and extracts
essential features like edges, textures, and patterns. These features are passed to the subsequent
layers for further analysis.
2. Neck: The neck enhances the features extracted by the backbone using structures like a Feature
Pyramid Network (FPN) or Path Aggregation Network (PAN), which help detect objects at
different scales. It combines both low-level and high-level features to improve detection accuracy
for objects of various sizes.
3. Head: The detection head is responsible for generating the final predictions. It outputs bounding
boxes, class probabilities, and object confidence scores for each detected object. YOLOv8 uses
anchor-free detection, which simplifies the detection process and improves speed.
4. Prediction: The final stage of the module outputs the objects detected in the image along with
their respective bounding boxes, confidence scores, and classes. The model balances between
high speed and accuracy, making it suitable for real-time applications.

YOLOv8’s object detection module is optimized for performance, employing efficient computation
techniques while maintaining high detection accuracy.

Telegram ChatBot

The Telegram chatbot module is designed to interact with users via the Telegram messaging platform,
automating tasks and providing information. Key components include:

1. Telegram Bot API: The bot communicates with Telegram’s servers using the Bot API, allowing
it to receive messages, commands, and send responses to users. It also enables bot
authentication and interaction with different Telegram features like groups, channels, and direct
chats.
2. Message Handler: This module listens for incoming messages or commands from users. Based
on predefined conditions or patterns (e.g., specific keywords or commands), it triggers
corresponding responses or actions. It processes both text and multimedia messages.
 14

3. Natural Language Processing (NLP): If the bot uses AI, this component helps interpret and
analyze user input using NLP libraries (like spaCy or NLTK), enabling the bot to understand
context, sentiment, or intent, making conversations more dynamic and user-friendly.
4. Response Module: Based on the input received, the response module generates the appropriate
reply. It can pull responses from predefined datasets, APIs, or dynamically generate replies
using machine learning models or decision trees.
5. External API Integration: This module allows the chatbot to interact with external services
and APIs (e.g., weather, news, or AI models) to fetch real-time information and deliver it to the
user.
6. Database: Stores user data, preferences, conversation history, or settings. This helps personalize
interactions and retain context for future conversations.Together, these components ensure the
Telegram chatbot can handle user queries, automate tasks, and provide information efficiently.

4. Implementation

Power BI:
Creating a dashboard in Power BI involves several steps. Here's a general procedure to help you get
started:
1. Install Power BI Desktop:
● Download and install Power BI Desktop from the Microsoft Power BI website.
2. Connect to Data Sources:
● Open Power BI Desktop.
● Click on Get Data in the Home ribbon to choose your data source (e.g., Excel, SQL Server, web,
etc.).
● Load the data into Power BI by selecting the appropriate file or server and following the prompts.
3. Transform and Clean Data (Optional):
● Use Power Query Editor to transform and clean your data if necessary.
● Click on Transform Data to open the Power Query Editor.
● Perform actions like filtering rows, changing data types, merging queries, etc.
● Click Close & Apply to apply the transformations.
4. Create Data Model:
● Define relationships between different tables if your data is coming from multiple sources.
● Click on Model to see and manage relationships.
● Drag and drop fields to establish relationships.
 15

5. Build Visualizations:
● Click on Report to go to the report view.
● Select a visualization type from the Visualization pane (e.g., bar chart, pie chart, map, etc.).
● Drag fields from the Fields pane to the visualization to populate it.
● Format the visualization using the options in the Visualization pane.
6. Arrange Visualizations on the Dashboard:
● Add multiple visualizations to your report page.
● Resize and arrange them to create a cohesive dashboard layout.
● Add text boxes, images, and shapes as needed to enhance the dashboard.
7. Add Interactivity:
● Use slicers and filters to make your dashboard interactive.
● Drag a field to the canvas to create a slicer or add filters to individual visualizations or the entire
page.
8. Publish the Dashboard:
● Save your Power BI Desktop file.
● Click on Publish in the Home ribbon.
● Sign in to your Power BI account.
● Choose a workspace to publish your report to the Power BI service.
9. Create and Share a Dashboard (Power BI Service):
● Go to the Power BI service (app.powerbi.com) and navigate to your report.
● Click on the Pin to a dashboard option for visualizations you want to add to a dashboard.
● Create a new dashboard or pin to an existing one.
● Share your dashboard with others by clicking on the Share button and entering the email
addresses of the recipients.
10. Set Up Dashboard Tiles (Optional):
● Customize the dashboard by rearranging and resizing the tiles.
● Set up alerts, drill-through actions, and other interactive features as needed.

Object Detection(Human) using YoloV8:

To get started, create a free Roboflow account.
After reviewing and accepting the terms of service, you will be asked to choose between one
of two plans: the Public Plan and the Starter Plan. Choose the public plan and enter a name to
the workspace. After click on create workspace.
 16

Then, you will be asked to invite collaborators to your workspace. These collaborators can help
you annotate images or manage the vision projects in your workspace.

Once you have invited people to your workspace (if you want to), you will be able to create a
project.
 17

For this example, we will be using a dataset of cars to train a model that can identify screws in
a kit. This model could be used for quality assurance in a manufacturing facility. With that said,
you can use any images you want to train a model.

Leave the project type as the default "Object Detection" option since our model will be
identifying specific objects and we want to know their location within the image.

Click “Create Project.” to continue.

Paste the youtube link and then click next it will process the video here the video is nothing
but a series of images here we will take 1 frame/second then click on choose frame rate.

Now click on universe and click on self-driving there are 100 of datasets select “vehicles
Computer Vision Project” click on download this dataset. In format select YOLOv8
 18

After selecting YOLOv8 click on continue. Copy the code now click on a notebook from your
model library select YOLOv8 after clicking YOLOv8 you can see Train on colab on the
rightside open train on colab. Here you can see the following codes.

Now run the cell. This cell is used to check whether you have connected to GPU(Graphics
Processing Unit) or not.

import os HOME = os.getcwd()

print(HOME)

Now run the above cell this cell works the current directory to home

# Pip install method (recommended) !pip install ultralytics==8.0.196

from IPython import display
display.clear_output()
import ultralytics
ultralytics.checks()

Run the above cell it will install the YOLOv8

# Git clone method (for development)

# %cd {HOME}
# !git clone github.com/ultralytics/ultralytics #
%cd {HOME}/ultralytics

# !pip install -e .
# from IPython import display
 19

# display.clear_output()

# import ultralytics

# ultralytics.checks()

You no need to execute the above cell

from ultralytics import YOLO

from IPython.display import display, Image

Run the above cell this cell will import yolo. After executing the four cells scroll down you
will find a cell in Preparing a custom dataset. After step 5 you can see a cell.

Now run the Custom Training cell

%cd {HOME}
!yolo task=detect mode=train model=yolov8s.pt
data={dataset.location}/data.yaml epochs=25 imgsz=800 plots=True

After running the above cell, you can upload the image or video so that it can label all the
objects. In runs-detect-train download the best.pt file this file is the model file
 20

Now run Inference with custom model

%cd {HOME}
!yolo task=detect mode=predict
model={HOME}/runs/detect/train/weights/best.pt conf=0.25

After running the above cell, the result will be saved in runs/detect/predict

import glob

from IPython.display import Image, display

for image_path in
glob.glob(f'{HOME}/runs/detect/predict3/*.jpg')[:3]:
display(Image(filename=image_path, width=600))

print("\n")

After running the above cell, the below output is generated

 21

Paste the video link in source After running the cell you will find the video with the detecting
of objects using YOLOv8

%cd {HOME}
!yolo task=detect mode=predict
model={HOME}/runs/detect/train/weights/best.pt conf=0.25
source=/content/test.mp4 save=True

After running the above cell you will find the result video in runs/detect/predict2

Connecting to google drive:

from google.colab import drive drive.mount('/content/drive')

After running the above cell you can find the result video in your google drive

Telegram ChatBot

Steps to Build a Chatbot

1. Define the Purpose
 22

● Determine what tasks the chatbot should perform (e.g., customer support, FAQ, booking system).

2. Choose a Platform

● Select a chatbot development platform that suits your needs (e.g., Dialogflow, Rasa).

3. Design the Conversation Flow

● Outline how the chatbot should interact with users.

● Create conversation scripts and decision trees for rule-based chatbots.

4. Implement NLP

● Train the chatbot to understand user inputs by defining intents (user goals) and entities (specific
data points).
● Use the chosen platform’s NLP tools to build the model.

5. Integrate Backend Services

● Connect the chatbot to databases, APIs, and other services to perform necessary actions (e.g.,
retrieving data, processing transactions).

6. Test the Chatbot

● Conduct thorough testing to ensure the chatbot handles various scenarios and edge cases.
● Use beta testing with real users to gather feedback.

7. Deploy the Chatbot

● Deploy the chatbot on the desired platform (e.g., website, messaging apps, voice assistants).
● Monitor performance and make iterative improvements based on user interactions.

Creating a new Telegram chatbot

Go to Telegram app, log in to your account, or create a new one. Type in @BotFather in the
search field, and go to this bot
 23

Click Start to activate the BotFather chatbot.

 24

You will receive a list of commands you can use to manage bots.
Select or type in the /newbot command, and send it.

Choose a name for your bot — your subscribers will see it during your conversations. You also
need to pick a username for your bot so that users can find it using search. Your bot username
must be unique and end with the word bot.

Once you choose a name for your chatbot, it will be created. You will receive a message with a
link to your bot (t.me/<bot_username>), recommendations on how to set up a profile picture,
description, and a list of commands you can use to manage your new bot.
 25

To connect to your bot you need to copy the token and place in TelegramBOT TOKEN
#AIMER Society - Indian Servers
!pip install pyTelegramBotAPI
!pip install openai
!pip install google-generativeai #For Google Gemini #AIMERS
!pip install anthropic
TelegramBOT_TOKEN = '7232921455:AAHPW7dW9q-UuMuyzDJQveya91UlK5dZyUM'

Now run the above cell.

Generate API key and paste in genai.configure and run the cell.

Latest version #Gemini API #AIMER Society #IndianServers import

telebot import os

import google.generativeai as genai

genai.configure(api_key="AIzaSyAe1eaPQmRKnLnLGDCKvzOxgS90oZEJy_I")
 26

# Set up the model

generation_config = {

"temperature": 1,

"top_p": 0.95,

"top_k": 0,

"max_output_tokens": 8192,

safety_settings = [

"category": "HARM_CATEGORY_HARASSMENT",

"threshold": "BLOCK_MEDIUM_AND_ABOVE"

},

"category": "HARM_CATEGORY_HATE_SPEECH",

"threshold": "BLOCK_MEDIUM_AND_ABOVE"

},

"category": "HARM_CATEGORY_SEXUALLY_EXPLICIT",

"threshold": "BLOCK_MEDIUM_AND_ABOVE"

},

{
"category": "HARM_CATEGORY_DANGEROUS_CONTENT",

"threshold": "BLOCK_MEDIUM_AND_ABOVE"

},
 27

model = genai.GenerativeModel(model_name="gemini-1.5-pro-latest",

generation_config=generation_config,
safety_settings=safety_settings)

convo = model.start_chat(history=[ ]) bot =

telebot.TeleBot(TelegramBOT_TOKEN)

@bot.message_handler(commands=['start', 'help']) def send_welcome(message):

bot.reply_to(message, "Welcome! The MOST POWERFUL AI BOT from

IndianServers")

@bot.message_handler(func=lambda message: True) def

handle_message(message): try :
print(message)
convo.send_message(message.text)
bot.reply_to(message, convo.last.text)

except Exception as e:
print(f"An error occurred: {e}") bot.reply_to(message,

"Sorry, I
couldn't process your request.")

Run the above cell to see the conversation between you and bot.
 28

5. Result Analysis

5.1 Power BI:

My Dashboard-
 29

5.2 Object Detection using YoloV8

Vehicle Detection Model Output:

 30

5.3 Telegram ChatBot

 31

6. Conclusion and Future Work

My artificial intelligence internship has been a truly transformative experience, offering me valuable
exposure to practical applications of my academic knowledge. Throughout the internship, I had the
opportunity to work on various projects, gaining hands-on experience with technologies such as chatbots,
generative AI, object detection, visual question answering, image classification, and computer vision. The
combination of guided tasks and a practical environment helped deepen my understanding of these critical
AI concepts.

As the field of artificial intelligence evolves, the demand for skilled professionals continues to rise across
different industries. Many organizations are looking to AI to drive innovation and improve their offerings.
With tools like Google Teachable Machine and DALL·E opening new possibilities, the career prospects
in AI are vast and exciting. This internship has provided me with the essential skills and knowledge needed
to succeed in this dynamic field and pursue a promising career in artificial intelligence.

Looking ahead, the rapid advancements in AI present endless opportunities for innovation and problem-
solving. The skills I’ve gained will allow me to develop intelligent solutions that enhance user experiences
and improve automation in various sectors. The future of AI holds great potential, and
I’m excited to be part of this evolving and impactful field, contributing to advancements that can benefit
society.